|
With the tremendous growth of sensitive and security-critical processing on embedded and pervasive platforms, the threat model for secure electronics is expanding from software into hardware. A wide range of fault attacks, based on physical manipulation of the electronics operating environment, is now available to the adversary.
The major outcome of this project is FAME, a collection of hardware techniques for microprocessor architectures to detect these fault injection attacks, and to mitigate fault analysis through an appropriate response in software. The FAME processor is developed both as an architecture concept as well as a chip prototype.
The FAME processor uses fault countermeasures that combine fault detection in microprocessor hardware with fault response in the software application. The fault detection in hardware uses static (design-time) and dynamic (runtime) techniques for in-situ fault detection. These fault-detecting hardware extensions are optimized for power and cost, and they can be enabled from the software application. This flexibility allows FAME to support non-critical applications at full microprocessor performance, while still offering full fault countermeasures for security-critical applications. The FAME processor chip demonstrates these techniques, as well as novel forms of fault analysis that are investigated in tandem with the development of FAME.
The impacts of this project are safer, more trustworthy microprocessors that are aware of their physical environment and the threats it poses to their internal processing. Such microprocessors offer the basis for cyber-security applications that can handle both logical as well as physical threats.
|
SaTC: STARSS: FAME: Fault-attack Awareness using Microprocessor Enhancements